High resolution (8 km) infrared (IR) GOES images encompassing the region of summer rainfall often termed the Mexican monsoon have been analyzed for the 10-year period from 1985 to 1994, using hourly images during the wet season (June - August) to estimate top cloud temperatures for various intervals (between -68C and 37C), and to create frequency images for each of them. The composites at the highest temperature were used to confirm the image navigation, using the difference in sea-land temperature during the afternoon, while the composites at cold temperatures have been used to describe the distribution of convective cloudiness.
After the frequency composites were prepared, the pixel values of IR temperature were extracted and compared with rainfall data, using a daily rainfall data set prepared for this region. This rainfall data set includes 150 climatological stations from the Mexican states of Sonora and Sinaloa. The rainfall data were analyzed by individual station and also by calculating averages over 1 by 1 degree regions (including at least 5 stations). Regression plots of rainfall versus percentage frequency of cloudiness for different temperature thresholds, and for different regions, will be presented. The interannual variation in cloudiness and rainfall will be described, and these variations will be related to the prevailing upper-tropospheric flow regimes during the corresponding months. From these comparisons, we will discuss the relative confidence that can be placed in GOES - IR based estimates of warm-season rainfall, especially for studies of climate variability over the region.